]> git.lyx.org Git - lyx.git/blobdiff - src/tex2lyx/text.cpp
Cmake build: support tests
[lyx.git] / src / tex2lyx / text.cpp
index db9f9b5adc7be4b10020a6b7453a2bbdcc40c6d7..7f7603c198b244057a8495edc381fd763e47db23 100644 (file)
@@ -130,17 +130,9 @@ const char * const supported_CJK_encodings[] = {
  * the same as supported_CJK_encodings with their corresponding LyX language name
  * please keep this in sync with supported_CJK_encodings line by line!
  */
-const char * const coded_supported_CJK_encodings[] = {
+const char * const supported_CJK_languages[] = {
 "japanese-cjk", "korean", "chinese-simplified", "chinese-traditional", 0};
 
-string CJK2lyx(string const & encoding)
-{
-       char const * const * where = is_known(encoding, supported_CJK_encodings);
-       if (where)
-               return coded_supported_CJK_encodings[where - supported_CJK_encodings];
-       return encoding;
-}
-
 /*!
  * natbib commands.
  * The starred forms are also known except for "citefullauthor",
@@ -822,8 +814,18 @@ void parse_box(Parser & p, ostream & os, unsigned outer_flags,
                        if (inner_type != "makebox") {
                                latex_height = p.getArg('[', ']');
                                translate_box_len(latex_height, height_value, height_unit, height_special);
-                       } else
-                               hor_pos = p.getArg('[', ']');
+                       } else {
+                               string const opt = p.getArg('[', ']');
+                               if (!opt.empty()) {
+                                       hor_pos = opt;
+                                       if (hor_pos != "l" && hor_pos != "c" &&
+                                           hor_pos != "r" && hor_pos != "s") {
+                                               cerr << "invalid hor_pos " << hor_pos
+                                                    << " for " << inner_type << endl;
+                                               hor_pos = "c";
+                                       }
+                               }
+                       }
 
                        if (p.hasOpt()) {
                                inner_pos = p.getArg('[', ']');
@@ -847,7 +849,7 @@ void parse_box(Parser & p, ostream & os, unsigned outer_flags,
                        if (!opt.empty()) {
                                hor_pos = opt;
                                if (hor_pos != "l" && hor_pos != "c" &&
-                                   hor_pos != "r") {
+                                   hor_pos != "r" && hor_pos != "s") {
                                        cerr << "invalid hor_pos " << hor_pos
                                             << " for " << outer_type << endl;
                                        hor_pos = "c";
@@ -1440,8 +1442,9 @@ void parse_environment(Parser & p, ostream & os, bool outer,
                // LyX doesn't support the second argument so if
                // this is used we need to output everything as ERT
                string const mapping = p.getArg('{', '}');
-               if ((!mapping.empty() && mapping != " ")
-                       || (!is_known(encoding, supported_CJK_encodings))) {
+               char const * const * const where =
+                       is_known(encoding, supported_CJK_encodings);
+               if ((!mapping.empty() && mapping != " ") || !where) {
                        parent_context.check_layout(os);
                        handle_ert(os, "\\begin{" + name + "}{" + encoding + "}{" + mapping + "}",
                                       parent_context);
@@ -1459,7 +1462,8 @@ void parse_environment(Parser & p, ostream & os, bool outer,
                        handle_ert(os, "\\end{" + name + "}",
                                       parent_context);
                } else {
-                       string const lang = CJK2lyx(encoding);
+                       string const lang =
+                               supported_CJK_languages[where - supported_CJK_encodings];
                        // store the language because we must reset it at the end
                        string const lang_old = parent_context.font.language;
                        parent_context.font.language = lang;
@@ -1842,28 +1846,31 @@ void fix_child_filename(string & name)
        if (!isabs)
                name = makeAbsPath(name, absMasterTeX).absFileName();
        bool copyfile = copyFiles();
-       // convert from absolute original path to "relative to master file"
-       string const rel = to_utf8(makeRelPath(from_utf8(name),
-                                              from_utf8(absMasterTeX)));
-       // Do not copy if the file is not in or below the directory of the
-       // master, since in this case the new path might be impossible to
-       // create. Example:
-       // absMasterTeX = "/foo/bar/"
-       // absMasterLyX = "/bar/"
-       // name = "/baz.eps" => new absolute name would be "/../baz.eps"
-       if (copyfile && rel.substr(0, 3) == "../")
-               copyfile = false;
        string const absParentLyX = getParentFilePath(false);
+       string abs = name;
        if (copyfile) {
+               // convert from absolute original path to "relative to master file"
+               string const rel = to_utf8(makeRelPath(from_utf8(name),
+                                                      from_utf8(absMasterTeX)));
                // re-interpret "relative to .tex file" as "relative to .lyx file"
                // (is different if the master .lyx file resides in a
                // different path than the master .tex file)
                string const absMasterLyX = getMasterFilePath(false);
-               name = makeAbsPath(rel, absMasterLyX).absFileName();
-               if (!isabs) {
+               abs = makeAbsPath(rel, absMasterLyX).absFileName();
+               // Do not copy if the new path is impossible to create. Example:
+               // absMasterTeX = "/foo/bar/"
+               // absMasterLyX = "/bar/"
+               // name = "/baz.eps" => new absolute name would be "/../baz.eps"
+               if (contains(name, "/../"))
+                       copyfile = false;
+       }
+       if (copyfile) {
+               if (isabs)
+                       name = abs;
+               else {
                        // convert from absolute original path to
                        // "relative to .lyx file"
-                       name = to_utf8(makeRelPath(from_utf8(name),
+                       name = to_utf8(makeRelPath(from_utf8(abs),
                                                   from_utf8(absParentLyX)));
                }
        }
@@ -1886,12 +1893,6 @@ void copy_file(FileName const & src, string dstname)
        else
                dst = makeAbsPath(dstname, absParent);
        string const absMaster = getMasterFilePath(false);
-       string const rel = to_utf8(makeRelPath(from_utf8(dst.absFileName()),
-                                              from_utf8(absMaster)));
-       // Do not copy if the file is not in or below the directory of the
-       // master (see above)
-       if (rel.substr(0, 3) == "../")
-               return;
        FileName const srcpath = src.onlyPath();
        FileName const dstpath = dst.onlyPath();
        if (equivalent(srcpath, dstpath))
@@ -2142,24 +2143,6 @@ void parse_text(Parser & p, ostream & os, unsigned flags, bool outer,
        while (p.good()) {
                Token const & t = p.get_token();
 
-       // it is impossible to determine the correct document language if CJK is used.
-       // Therefore write a note at the beginning of the document
-       if (have_CJK) {
-               context.check_layout(os);
-               begin_inset(os, "Note Note\n");
-               os << "status open\n\\begin_layout Plain Layout\n"
-                  << "\\series bold\n"
-                  << "Important information:\n"
-                  << "\\end_layout\n\n"
-                  << "\\begin_layout Plain Layout\n"
-                  << "This document contains text in Chinese, Japanese or Korean.\n"
-                  << " It was therefore impossible for tex2lyx to set the correct document langue for your document."
-                  << " Please set the language manually in the document settings.\n"
-                  << "\\end_layout\n";
-               end_inset(os);
-               have_CJK = false;
-       }
-
        // it is impossible to determine the correct encoding for non-CJK Japanese.
        // Therefore write a note at the beginning of the document
        if (is_nonCJKJapanese) {
@@ -2739,8 +2722,7 @@ void parse_text(Parser & p, ostream & os, unsigned flags, bool outer,
                        context.check_layout(os);
                        p.skip_spaces();
                        begin_inset(os, "Caption\n");
-                       Context newcontext(true, context.textclass);
-                       newcontext.font = context.font;
+                       Context newcontext(true, context.textclass, 0, 0, context.font);
                        newcontext.check_layout(os);
                        if (p.next_token().cat() != catEscape &&
                            p.next_token().character() == '[') {
@@ -2790,8 +2772,8 @@ void parse_text(Parser & p, ostream & os, unsigned flags, bool outer,
                                        os << "\n\\begin_layout Plain Layout";
                                        p.skip_spaces();
                                        begin_inset(os, "Caption\n");
-                                       Context newcontext(true, context.textclass);
-                                       newcontext.font = context.font;
+                                       Context newcontext(true, context.textclass,
+                                                          0, 0, context.font);
                                        newcontext.check_layout(os);
                                        os << caption << "\n";
                                        newcontext.check_end_layout(os);
@@ -3210,8 +3192,8 @@ void parse_text(Parser & p, ostream & os, unsigned flags, bool outer,
 
                else if (t.cs() == "href") {
                        context.check_layout(os);
-                       string target = p.getArg('{', '}');
-                       string name = p.getArg('{', '}');
+                       string target = convert_command_inset_arg(p.verbatim_item());
+                       string name = convert_command_inset_arg(p.verbatim_item());
                        string type;
                        size_t i = target.find(':');
                        if (i != string::npos) {
@@ -3729,7 +3711,11 @@ void parse_text(Parser & p, ostream & os, unsigned flags, bool outer,
                else if (t.cs() == "verb") {
                        context.check_layout(os);
                        char const delimiter = p.next_token().character();
-                       string const arg = p.getArg(delimiter, delimiter);
+                       // \verb is special: The usual escaping rules do not
+                       // apply, e.g. "\verb+\+" is valid and denotes a single
+                       // backslash (bug #4468). Therefore we do not allow
+                       // escaping in getArg().
+                       string const arg = p.getArg(delimiter, delimiter, false);
                        ostringstream oss;
                        oss << "\\verb" << delimiter << arg << delimiter;
                        handle_ert(os, oss.str(), context);
@@ -4423,6 +4409,79 @@ void parse_text(Parser & p, ostream & os, unsigned flags, bool outer,
        }
 }
 
+
+string guessLanguage(Parser & p, string const & lang)
+{
+       typedef std::map<std::string, size_t> LangMap;
+       // map from language names to number of characters
+       LangMap used;
+       used[lang] = 0;
+       for (char const * const * i = supported_CJK_languages; *i; i++)
+               used[string(*i)] = 0;
+
+       while (p.good()) {
+               Token const t = p.get_token();
+               // comments are not counted for any language
+               if (t.cat() == catComment)
+                       continue;
+               // commands are not counted as well, but we need to detect
+               // \begin{CJK} and switch encoding if needed
+               if (t.cat() == catEscape) {
+                       if (t.cs() == "inputencoding") {
+                               string const enc = subst(p.verbatim_item(), "\n", " ");
+                               p.setEncoding(enc);
+                               continue;
+                       }
+                       if (t.cs() != "begin")
+                               continue;
+               } else {
+                       // Non-CJK content is counted for lang.
+                       // We do not care about the real language here:
+                       // If we have more non-CJK contents than CJK contents,
+                       // we simply use the language that was specified as
+                       // babel main language.
+                       used[lang] += t.asInput().length();
+                       continue;
+               }
+               // Now we are starting an environment
+               p.pushPosition();
+               string const name = p.getArg('{', '}');
+               if (name != "CJK") {
+                       p.popPosition();
+                       continue;
+               }
+               // It is a CJK environment
+               p.popPosition();
+               /* name = */ p.getArg('{', '}');
+               string const encoding = p.getArg('{', '}');
+               /* mapping = */ p.getArg('{', '}');
+               string const encoding_old = p.getEncoding();
+               char const * const * const where =
+                       is_known(encoding, supported_CJK_encodings);
+               if (where)
+                       p.setEncoding(encoding);
+               else
+                       p.setEncoding("utf8");
+               string const text = p.verbatimEnvironment("CJK");
+               p.setEncoding(encoding_old);
+               p.skip_spaces();
+               if (!where) {
+                       // ignore contents in unknown CJK encoding
+                       continue;
+               }
+               // the language of the text
+               string const cjk =
+                       supported_CJK_languages[where - supported_CJK_encodings];
+               used[cjk] += text.length();
+       }
+       LangMap::const_iterator use = used.begin();
+       for (LangMap::const_iterator it = used.begin(); it != used.end(); ++it) {
+               if (it->second > use->second)
+                       use = it;
+       }
+       return use->first;
+}
+
 // }])